Journal article, 2013

In this paper, multi-group microscopic cross-section uncertainties are propagated through the DRAGON (Version 4.05) lattice code in order to perform uncertainty analysis on k(infinity) and 2-group homogenized macroscopic cross-sections. The test case corresponds to a 17 x 17 PWR fuel assembly segment without poison at full power conditions. A statistical methodology is employed for such purposes, where cross-sections of certain isotopes of various elements belonging to the 172 groups DRAGLIB library format, are considered as normal random variables. This library was based on JENDL-4 data, because JENDL-4 contains a large amount of isotopic covariance matrices among the different major nuclear data libraries. Thus, multi-group uncertainty was computed for the different isotopic reactions by means of ERRORRJ. The preferred sampling strategy for the current study corresponds to the quasi-random Latin Hypercube Sampling (LHS). This technique allows a much better coverage of the input uncertainties than simple random sampling (SRS) because it densely stratifies across the range of each input probability distribution. In order to prove this, the uncertain input space was re-sampled 10 times, and it is shown that the variability of the replicated mean of the different k(infinity) samples is much less for the LHS case, than for the SRS case. The uncertainty assessment of the output space should be based on the theory of non-parametric multivariate tolerance limits, due to the fact that k(infinity) and some of the macroscopic cross-sections are correlated. Therefore, for 10 replicated samples each containing 100 elements, the total output sample is composed by 1000 calculations. This sample size is more than enough to infer that the multivariate output population is covered 95% with a 95% of confidence. On the other hand, statistical sensitivity analysis was performed in order to know which microscopic cross-section has the greatest impact on k(infinity) predictions. It was found that the fission cross-section of Uranium 235 is the dominant input parameter for this particular case, because the computed JENDL-4 variances for such reaction are very high at thermal and resonant regions compared to other variances that for instance, can be computed based on other nuclear libraries such as ENDF/B-VII.1

JENDL-4 covariance data

DRAGONv4.05 code

Latin Hypercube Sampling

Statistical uncertainty analysis

Chalmers, Chemical and Biological Engineering, Nuclear Chemistry

Chalmers, Applied Physics, Nuclear Engineering

Chalmers, Chemical and Biological Engineering, Nuclear Chemistry

0306-4549 (ISSN)

Vol. 57 230-245Other Engineering and Technologies not elsewhere specified

Energy

10.1016/j.anucene.2013.01.061